Artificial Neural Networks 0909.560.01/0909.454.01 Fall 2004

1
Artificial Neural Networks0909.560.01/0909.454.01
Fall 2004
Lecture 4October 4, 2004

• Shreekanth Mandayam
• ECE Department
• Rowan University
• http//engineering.rowan.edu/shreek/spring04/ann/

2
Plan

• Multilayer Perceptron
• Recall Architecture
• Signal Flow
• Learning rule - Backpropagation
• Final Project Assignments
• Lab Project 2

3
Multilayer Perceptron (MLP) Architecture
4
MLP Signal Flow
j
j
j

• Computations at each node, j
• Neuron output, yj
• Gradient vector, dE/dwji

5
Backpropagation
Notation
At a node j,
dj(n)
vj(n)
yj(n)
wji(n)
yi(n)
ej(n)
j(.)
-1
6
Backprop. (contd)
Notation
If node j is a hidden node,
dk(n)
vk(n)
yj(n)
vj(n)
yk(n)
wji(n)
wkj(n)
yi(n)
ek(n)
j(.)
-1
j(.)
7
MLP Training
k
j
i
Right
Left

• Forward Pass
• Fix wji(n)
• Compute yj(n)
• Backward Pass
• Calculate dj(n)
• Update weights wji(n1)

y
x
k
i
j
Right
Left
8
MLPs in Matlab
http//engineering.rowan.edu/shreek/spring04/ann/
demos/mlp.m
9
Lab Project 2

• http//engineering.rowan.edu/shreek/spring04/ann/
  lab2.html
• UCI Machine Learning Repository
• http//www.ics.uci.edu/mlearn/MLRepository.html
• Face Recognition Generate images

10
Summary